 Good afternoon, thank you very much for making the time to hear some of these thoughts on what, as you've heard, is a set of interesting and controversial issues. My talk addresses a number of examples. There are plenty of questions. There are not that many answers, but I hope in our discussion we'll have a chance to work through some of those points in a little more detail. Mae gweithgannaeth bod ydy'r gwstiannod hefyd wedi cydnabod i neud y mynd i'r amser. Yr ydych chi dde un yn amlwg, drwy'r Gweithgai a'r Cymru, Aledinogion, unrhaed, dyma'r Unigach gyda Izan Harrel, ac yn llwyso ar gyfer yma yw'r sgwrdd hefyd o'u ddraen o'u cydnabod ar yr Asgol H ôl Llyw yn Torronto Harry Arthur's Fellowship, that's Harry Arthur's, not me in my time machine. And in that context, engaging with some of the debates regarding technological development in Canada, which I'll talk about today. So the idea for this talk is to look at some of the controversial aspects of smart cities and how that has engaged legal and policy debate. And to do so by examining a number of areas. I start by talking about city developments and focusing on a project in Toronto. I'll continue by exploring the development of social credit systems in China. After that, I'll identify what I think is an emerging area where these questions of law and technology are being debated, and that's in relation to mobility and mobility data in particular. That's what we talk about scooters. And then I'll finish with a few words around faith and recognition and outline a couple of strategies, almost coping strategies are again rather than answers. But ways that might frame the discussion that we will have in the time we have remaining at the end of the talk. So smart, who doesn't want to live in a smart city? As explained in one guardian piece of a few years ago, you'll wake up. Everything's going to be great. You will have an energy efficient house. Your electric car is going to be outside. The sensors will guide you to your workplace. And everybody shall have a good life. And this is how we were told it was going to be. Even though we might sometimes wonder looking at 20th century science fiction when the jet pack is going to show up, we see a set of aspirations around making urban life in particular safer, more efficient, have less of an impact on climate and so forth. And so it's not surprising that many have adopted strategies or goals in this regard. I've just given one example there from the Scottish digital strategy that we will have the smart city when we use digital technologies to manage our resources and our infrastructure in a sustainable way, more efficient greener places to live and do business, et cetera. Or as a famous essay by one of the founders of Airbnb put it, our cities have begun to feel mass produced. We live closer together, but we drift apart. And so in the Airbnb model, we think of how we use technology to change the way in which we distribute, share, utilize resources. And what I intend to do in the next moment or two is just focus on one area where a city has tried to develop a smart project or a smart district and how the reactions in that project tell us something quite significant. And that's this patch of land, a 12 acres on the waterfront of Toronto or as the global mail put it, 12 acres of unsightly, soggy Toronto waterfront real estate, which some of you may know has been the subject of a long running set of debates. Involving the agency, which is responsible for work front redevelopment work front Toronto, which brings together federal, provincial and municipal levels of government with sidewalk, which is a company within the alphabet or Google family. And the process of acquiring this land and developing a strategy for how sidewalk would build a sort of a model of the city of the future. Has been controversial. There are a number of key documents here. One of them is over a thousand pages long. So I look forward to reading it from this podium in detail and we shall we shall take a break, maybe about six o'clock. We'll be fine. No, the we can understand sidewalks aspirations here. And we look at the submission statement, reimagining cities to improve quality of life. Although as the Canadian public broadcasted the CBC put it, welcome to the neighborhood. Have you read the terms of service? And it's that idea that on one hand, sidewalk would would build an area that is clean, efficient, green, a pleasure to live in, whether the price in terms of the collection of of data, personal or otherwise, the quality might mean for different levels of government and how they relate to a powerful global corporation. And even in negotiating this are are quite significant. But of course, it is that it is that data and technological dimension that characterises what sidewalk is trying to do. Model cities and dream cities are not new. This is the Minnesota experimental city from the 1960s, subject of a brilliant documentary a couple of of of years ago. And we see this idea that, you know, if you are starting from scratch, you can you can build at different levels and you can actually anticipate all the the different needs. We look at some sidewalks planning documents as one of the early documents that went to to work from Toronto. We see them imagining something very similar infrastructure, different levels, but a digital layer upon which all of the all the exciting stuff might might happen. And so it wasn't surprising that at least in its earlier public documents, sidewalk described their plans as to build the most measurable community in the world, a former colleague at Newcastle University once described, Bob Hollins once described, the ubiquitous city that when you embed sensors for data collection in a city that gives you opportunities, but also particular threats, particularly when you have this level of embedding or building in the MIT technology review in its list of breakthrough technologies last year. Said the sensing city was one of them and sidewalk was the the paradigmatic example of that. One of the remarkable things about sidewalks response to controversy is how it has pivoted from talking about the most measurable community in the world to focusing on the quality of life benefits. And so we see great emphasis now, not just on reassurances around privacy, but also pointing to why this digital layer would exist. You can really get to the heart of a Canadian, for instance, by telling them that there will be tiles and the tiles will know when the weather is going to change and therefore the will melt the snow quicker and so on. You can see that kind of the idea that a conversation about making a measurable community does not necessarily give you the positive feedback, whereas a conversation about a city where you never slip on the snow might give you a different sort of reaction. Sidewalk has responded in in other ways. And actually, even if the project never goes ahead, what we have learned from it is some of the tools that we might need for projects of this nature. So this is one of them. This is the the work that sidewalks currently doing on design transparency in the public realm. And it's the idea of coming up with a vocabulary of symbols for sensors and for data collection in in common areas where you would be able to communicate what could be quite complex concepts in a quick and relatively easy fashion. Similarly, sidewalk have experimented with but have now pulled back from thinking about an urban data trust as a way of bridging some of those public and private sector dimensions and having a governance framework for the type of data that would be collected by a ubiquitous or a sensing city. Most recently, we have seen the negotiations between the public authorities and sidewalk run into a little bit of trouble. One of the advisory groups set up to feed into this process described what sidewalk we're proposing as tech for tech's sake. And the various commitments have been made in relation to personal data, facial recognition and so on, which I will come back to. And at this stage, it still boils back in the court of work from Toronto as to whether they would give approval, which will be necessary because of the ownership of land for the project to to go ahead. But we have seen what happens when a project develops and indeed the range of issues that need to be thought about, particularly where there is this kind of formal process for approval that's that's taking place. And from the point of view of a of a public authority, of course, there are a number of different goals here. This is not in the first instance about inventing a particular form of sensor or coming up with a way of collecting data. This would address broader public policy goals, which in the case of Waterfront Toronto, included economic development, sustainability and mobility. And so I think those that are pitching smart city solutions are increasingly seeing cities and public authorities trying to articulate what they want out of it and how this would contribute, for instance, to new economic or social development goals. And that's an interesting inversion because in the earlier wave of smart city projects it was often an off the shelf solution that was pitched. And then the city, well, do you want it, yes or no? Whereas now we're actually seeing an articulation of needs and quite a number of cities maybe that are later to the picture in terms of smart city development are having that second mover advantage of being able to see that having a sense of what you want is a better position for negotiations rather than simply looking at what's being pitched at the major conferences and markets. And the data question is right at the heart of it. This is a PWC document published late last year identifying the different things that you would want to know about before starting a smart city project, including what type of data, how it's going to be collected, is there a monetisation strategy, answer, yes, and so on. So we see not a disavowal of the commercial dimension because that's always been a part of it, but a recognition that that's part of what needs to be debated and indeed to be at least resolved or to have a structure for resolution. Well, I move from there to something slightly more controversial. So maybe the ultimate in smart city development whether you like that or not. And that's some of the work that's been developing in Chinese cities regarding what's sometimes described as a social credit system. Also, as I will say, it's more like a set of social credit systems. And here I draw upon work that has been recently published, which I've co-authored with my friend and colleague, Mattias Seams. And in this work, we try to understand the significance of rating and reputational systems, including from a legal point of view. This is something again that we understand from fictional depictions, minority report and beyond, but we also see reported as a threat as in this particular headline, this dystopian technology could be contagious. And what we tried to do there was understand how what was being developed in China that would purport to measure sincerity, honesty and integrity could actually be explored and would offer us a valuable set of lessons regarding technological change and how law interacts with that. And we argued that in China there are at least three different versions of what is being reported as social credit, one being a very top-down approach. Perhaps the closest version of that would be lists of defaulters, those who have unpaid fines or have not cooperated with courts, although that has been expanded. The second example being pilot cities, which is much more closely tied to surveillance and often involves a scoring rationale where good behaviour is rewarded and bad behaviour is punished through changes to scores. And the third being what has primarily developed in the Chinese tech sector through Alibaba and Tencent of coming up with different ways and often quite novel ways of doing what credit scoring has conventionally done. What we're also doing, of course, is, again, coming back to the fictional dimension, talking about Black Mirror and a famous episode called Nosedive where we imagined a world where everyone is rated on every interaction and that your score then affects your ability to travel, to get access to services, and ultimately to remain free. While we were doing the project, we were kind of kept account on how many times Black Mirror and social credit appeared in the same news story and we gave up after a few thousand. But again, it taps into that anxiety of the nearly there technology and how we might respond. In our work, we argue that we have been doing rating and reputations for a long time. This has been facilitated by technological change and by legal measures present or absent, as well as by commercial practices. We trace a thread from the development of credit rating systems in the 19th century US, particularly at the frontier, whereas capitalism rolled further west. The idea of working out whether someone was going to pay you back and not relying on where did they go to school or did they have a good family background actually trying to think of an objective and reproducible measurement of creditworthiness was so important to US developments in the 19th century and of course then computerized in the 20th. We talked about eBay and how ratings contributed to the success of eBay and indeed the willingness to engage in repeat purchasing. And of course the sharing economy, which in some ways is what the Black Mirror episode is trying to get at. You're already rating your stay in an Airbnb and both the driver and the passenger are rating each other an Uber. So let's just take that to the next level. Perhaps not to the level of, as you see on the slide, credit score dating, where it all starts with a number. But we have seen the development of alternative models for the unbanked, for less robust financial systems, where using other signals might actually supplement and ultimately outgrow conventional approaches to credit. What we did in that work and how we tried to deal with that a long history alongside the Chinese developments was to understand that even systems that appeared to be working in different political or economic contexts would still have things in common. So we looked for instance at who was responsible for a system where you could choose to take part in it or were required to. And of course that wasn't always certain. So one would say, well you have a free choice as to whether you use a service like Uber or Lyft. And you do until there's nothing else available. Or indeed the conventional jobs in the sector start to dry up and therefore rating systems become more essential. And indeed, applying the framework that we developed in our research, we can see that every choice whether made by a pilot project in China or a private company in the US has consequences. So if you give a user a choice on whether to opt in to a system that can be games, that can be manipulated and players with bad scores can simply withdraw from that market. On the other hand, if you have no choice in taking part at all then the actual interests of the users may be disregarded. Similarly, if you have a plethora of systems which are developed independently, your reputation is not portable which creates a lock-in effect. On the other hand, if everything is put together then the fact that you have been a bad driver might start to affect access to other services and so on. So each situation has consequences and we need not be complacent about this because we already see this type of phenomenon happening. We see for instance penalty points developed in a formal legislative context being picked on by insurance or in car hire. We see something like the football banning order under English legislation which deals with a complex set of public and private activities in order to achieve a law enforcement goal. But again thinking of blacklists in a way. Not of course what is happening in some of the Chinese experiments but at least having some structural similarities and therefore it's often the safeguards and the detail that tells them apart. And the legislative approaches here of course can be interesting. So in our work we explored how the GDPR would restrict the development of some of the models for seeing in China in this jurisdiction and others. But we also see of course other legal areas including equality law having a role to play as well as the clear recognition for instance by consumer protection authorities that there is a place for ratings. Particularly where there is legal uncertainty or where there is a limited role for a conventionally constructed piece of consumer legislation. The kind of transparency of ratings and the kind of peer-to-peer dimension has certainly been relied upon tacitly as a way of providing for strong outcomes. And so in our work here we see that the incorporation of rating and reputational technologies into the urban environment is something that is developing at a very fast pace in a number of Chinese cities and potentially linked to a number of other developments there. And the question that many cities here will have to face is not just what is lawful in terms of data protection at all but also whether there is a case for incorporating rating and reputational systems particularly where the hardware to do so may already be under development and capable of application on the Chinese streets as much as it is on the streets of Dublin. I'll move now, if I may, to a slightly different version of the same problem. And that's the question of how to handle data not in relation to individuals but to their movement and the devices that they use. And I'll start by talking about scooters, as I promised. This has been an area where we have seen very rapid development and we'll probably see it move on. And if they gave this talk in two years' time, it would have to put a picture and explain what an electric scooter was. But for instance in the United States, the number of trips on scooter-sharing systems being the same last year as the number of trips on fixed bike-share schemes. And what we see there is a combination of a number of technologies such as GPS apps and so on. But we also see fulfilling a real need as to dealing with some of the gaps in existing public transit systems. And the sort of experimental nature of doing something new, particularly where the regulatory context is, is unsure. What's interesting, though, in the United States is that many cities have learned a lesson from the sharing economy and actually want to think about the relationship with the companies now rather than two years after everybody's using the services. And so drawing upon some particular aspects of US municipal law where there's a permitting framework for putting furniture in a public space or there are ways in which cities can act quite quickly to adopt bylaws, many cities have adopted a licensing or a permit system where if you are Lyme or Bird or one of the operators here, you need a licence. And once you're into that licensing arrangement, then you start to attach conditions. And this is important because, as I say, this won't just be about scooters. This is the way in which we start to work out the relationship between public authorities and those that are developing different types of autonomous vehicles. This is about the major players in the sharing or gig economy starting to think about where they go next, including in some cases just buying bike and scooter companies. We see it even in relation to the biggest crisis in e-commerce, which is parcels. The number of parcels delivered and what that means for city traffic and how you might think about whether there's a technological solution to that drones or otherwise. All of this, actually the scooter or mobility arguments give us a chance to test it out. And we've seen this, for instance, in debates in the UK around having a national mobility strategy and then trying to take what has been argued to be a joined-up approach. And I come to you today from, as this headline tells you, the most car-dependent city in the United Kingdom. But my example here is from another car-dependent city, which is Los Angeles. What Los Angeles has tried to do is develop a strategy for mobility data. And why would it want to do this? Well, the data that flows, for instance, from a company operating a scooter-share scheme or a bike-share scheme is useful in all sorts of ways. If a city has a set of rules, then it would want to see that those rules are being enforced. If a city can understand who is using scooters and where they're being used, then that tells you a story about conflict over use of cycle lanes. It tells you where it's busy. You can use some of that data to build the Holy Grail of Transport, which is the multimodal trip planner. How can I get from here to there with all of the available options at this particular time right now in the device that I'm holding for our research uses too? Many cities, particularly those that have had long-standing issues with gridlock, imagine the idea that not just micro-availity devices, both autonomous vehicles and so on, if you have a level of control, the idea of the city traffic controller as air traffic controller where you can stop, reroute, deal with crisis and so on, as well as, of course, the long-standing questions around how this data will be useful from a law enforcement or advertising point of view. So what Los Angeles did, very interestingly, was develop a specification for data and then tried to use the permitting process to require operators to provide data in that form. Initially just from a compliance point of view, although as adopted in a number of other cities, also with that sort of reverse flow air traffic control rationale, and having outgrown its Los Angeles roots, was recently converted into a new organisation, the Open Mobility Foundation, which seeks to bring together public and private stakeholders to govern this controversial area of mobility data. Not everybody is pleased, as you might imagine. There have been attempts to deal with this at a legislative level, which in case of Los Angeles then was one level up, and the State of California are trying to set some restrictions on what data cities could require, and Uber and a number of players are arguing that this is a violation of state privacy law to require it in this way. What we've seen here is, on one hand, an attempt to get ahead of the game by understanding what data would be valuable and how you would build a framework for it. Governance, but also then how that can itself become the story, because the story of the mobility specification is probably bigger than the scooters itself, and certainly will outlive it, because the compromises or the balances that are struck are a warm-up for relationship, for instance, between public authorities and developers of autonomous vehicles. Using that as a flash point, I want to move on to one more, which is facial recognition. It is an area where I think facial recognition is part of the smart city story, and it has been integral to the success of a number of the social credit projects in China, but it also raises some quite similar issues to what we were talking about a moment ago regarding mobility. Sidewalk incidentally has ruled out facial recognition in order to sort of shore up its public support for what it's trying to do. Some of you would have seen the New York Times report from a couple of days ago regarding one of the private sector projects, Clearview AI, which, as it says, has scraped a few billion images from social media and other sources to develop an app, which is proving very popular with law enforcement in the United States, because it apparently vastly exceeds the capacity of the systems that FBI and local enforcement actually have. This has long been a case of warnings in academic literature about whether this will be happening in China, whether it will be happening in the private sector, and this has certainly brought it into focus. In the United Kingdom, the Information Commissioner, as a number of you may recall, has recently issued an opinion on facial recognition technology by the police in public places. A certain amount of this is an analysis of lawfulness because this is processing of special category data, it's biometric, but also interestingly, the commissioner in the UK is pointing out that the current combination of laws, codes, et cetera, is not a robust regulatory regime for these technologies and therefore advises a statutory code of practice, not dissimilar to what was already created under the conservative Lib Dem government in the UK of 10 years ago now of a specialist commissioner for surveillance cameras and then for biometrics. In this jurisdiction, of course, the Data Protection Commissioner has issued guidance on body-worn cameras, which interestingly picks up the point that we kind of have a spectrum, starts from a conventional CCTV to body-worn cameras, and if you start to combine that with facial recognition technology or microphones over to points, the concerns are more significant. This has come to a head quite quickly in that the courts of England and Wales have already heard a challenge to the use of facial recognition technologies by the police in South Wales. This is a test case, in a way, because South Wales police are the national lead for the trials and had national funding to carry out these trials. The courts found that the particular project was lawful. It was permitted, in part, by the existence of common law police powers, which satisfied the various tests under Article 8 of the European Convention on Human Rights and, in a related way, what was required of data protection law. But the courts did accept against what the police had said that data protection law was engaged, that the European Convention on Human Rights was engaged, and the ICO itself says that this decision shouldn't be treated as blanket authorisation for the use of law of facial recognition technologies because certain features were part of this system, as design would not be present in order, so it wasn't the case that facial recognition was presumptively lawful in the United Kingdom. And it's certainly the court decision, which is under appeal, has provoked a good debate. The legislative debate, in the meantime, is moving on. It's not accurate to say that the UK is just about to ban facial recognition technology. There is a private members bill introduced in the House of Lords. That will certainly prompt a debate, but that in itself is not legislation. Also, the argument, of course, that facial recognition is entirely unregulated. Again, I think the court proceedings has demonstrated that perhaps there might be arguments for and against the impact of existing legislation or indeed human rights measures. But there are a certain number of things there. And we see it in the United States as well, famously in San Francisco, acting to ban facial recognition technology, followed by a number of other cities in the United States. This has been particularly linked with policing and racial justice. And it is starting to have an effect on smart city strategies, that the development, for instance, of sensors, cameras, the kind of technological infrastructure to deliver what might be an existing smart city strategy, is quite rightly being debated also in terms of race and communities. It's not easy to get a rule on this, right, though? San Francisco, unfortunately, had to go back and amend its rule because it seemed accidentally to prohibit staff from using iPhones, which sort of face ID on lock. On the other hand, Portland is debating whether not just to ban city use of facial recognition technologies, but also to restrict private sector use within the city, which raises interesting jurisdictional questions. There is a congressional hearing on the topic in the United States quite recently, the third in a year. We're again seeing a question of a temporary ban or a moratorium being put up with one member of Congress saying this is some real black mirror stuff we're seeing here. So again, we're seeing kind of the adoption cycle between science fiction and let us debate shortening as we go. And really, just as we prepared for this event today, we saw an interesting leak of draft paper from the European Commission regarding the broader topic of AI and whether legal change might be necessary. And this echoes some work that has already emanated, not just from the Commission, but from the EU Fundamental Rights Agency, some of which was public. Now, this is some distance away from being a formal legal proposal. And even the Commission itself in the full leak document doesn't seem entirely convinced that the case has been made out for a restriction or a moratorium. But the wide reporting of the document is really significant and it indicates the level of interest in the topic. Again, as a number of better qualified commentators have pointed out, we do run the risk of focusing all of our scrutiny on facial recognition rather than on surveillance or the wider aspects of governance and control in the smart city. But even again, in this jurisdiction, the reaction to the plans of the National Children's Hospital reported before Christmas, where that Chinese link was emphasised and often in news reporting, this is what is happening in Chinese cities. Do you want it to happen in your hospital? And the very recent reporting regarding the plans of a school to use a version of facial recognition technology and very quickly to say that it wasn't going to do so again. We will see, for instance, and this is one thing that we've learned from the South Wales case, the infrastructure of data protection law starting to kick in here in terms of impact assessments, in terms of what's required under UK legislation because it's the law enforcement directive of having a policy, which is a specific statutory requirement in the UK, starting to have an impact on how we debate this. So, to spend the last moment or so, just reflecting on what we've addressed in this talk and then handed over to you. We see across all of the examples I've talked about today what James Bridal called the presentation of new technologies as inherently emancipatory, this type of computational thinking where, as in the report on sidewalk, that we're doing tech for tech sake and that we think that this is the way to go. It's interesting to see this right now because as many others have said, we are at a point in a cycle where there is broader criticism of the power of tech companies in a so-called tech clash and how dealing with something like facial recognition or mobility data right now is different to how we might have dealt with it five years ago when perhaps we might have been more tolerant of this. To borrow a question that emerged from a fantastic multi-year project on smart cities at Manut University, co-ordinated by Professor Rob Kitchen, ask questions like what would the smart city look like if it was reconfigured by civic infrastructure, by which we mean way methods for participation, for debating the costs and benefits of the type of change that we have outlined today. And in many cases, as I've argued elsewhere in earlier research, this is not just about the role of public authorities because many of the most controversial issues in the smart city are at that intersection of public and private space where we see different approaches and different legal doctrines being relevant, ranging from contract to property and beyond. So I pause here really just to mention the Jetsons. Apologies to those that have heard me do this before. The Jetsons are a version of the future. That's very much in the past. But there's recently been a reboot of the Jetsons story. So the original version, of course, was in a way was how great life would be in the future. But the reboot, which is in graphic novel four and tells a backstory. And in a new version, the reason why the Jetsons ended up where they were was that the wealthiest of the earth-built space stations has protection against the environmental apocalypse. And they got out just in time, but everybody else didn't. So they ended up in the life of the future because everybody else died. This is deliberately provocative, but it reminds us that technological change doesn't always make life better that the benefits of tech aren't even distributed and even a utopian story or a childhood memory has dystopian underbelly. Now today I haven't told you when the flying car is going to arrive or indeed how it might help you the next time the tractors block the streets. But what I've tried to show is that the four areas we've talked about, Smart City, social credit, scooters and mobility and surveillance, we see these interconnections, the importance of facial recognition in enforcing social credit, the link between mobility data and achieving the environmental and efficiency goals of the Smart City. And we see recurring themes, search for data governance models, the concern about tech for tech sake, the blurring of lines between private and public and a need to articulate which social or economic goals are being pursued in these major projects and how that relates to trust and explanation. Whether it be Toronto, Shenzhen, Los Angeles or South Wales, what we're trying to do is imagine the city of the future and how long I play a role in doing that, how I might unlock and build trust. As National Geographic reminded us in a special issue last year, two-thirds of the world's population will live in a city by 2050, yet we still struggle with some of these issues of planning and regulating urban life and they have been cut off from the debate around law and technology for far too long. The image with which I've opened and closed is Walt Disney's original sketch for the experimental prototype community of tomorrow. Never built although it lives on in the name of the Epcot theme park. The controversies that we're talking about today challenge us to consider what are the legal measures we need for the community of tomorrow as we start to imagine the better life, the dream city and so I look forward to the discussion that we'll have in our remaining time. Thank you very much.